On 9/23/2015 8:29 AM, Sam Hartman wrote: > I tend to agree with John and others who have > suggested the document should be more clear about its assumptions. Assumptions, models, details, etc., with operational and risk implications. The document has a number of places that need considerable elaboration of these. In the case of the trust model, the document is, apparently, introducing an entirely new 'model'. As has been noted, the challenge of local-part handling is another, surprisingly-basic component to the mechanism that needs careful -- and complete -- handling in the spec, if the spec is to be a full protocol specification. And so on. d/ -- Dave Crocker Brandenburg InternetWorking bbiw.net